Sympathetic and Baroreceptor Reflex Function in Neurally Mediated Syncope Evoked by Tilt

Sympathetic and baroreceptor reflex function in neurally mediated syncope evoked by tilt.

R Mosqueda-Garcia, … , R M Robertson, D Robertson

J Clin Invest. 1997;99(11):2736-2744. https://doi.org/10.1172/JCI119463.

Research Article

The pathophysiology of neurally mediated syncope is poorly understood. It has been widely assumed that excessive sympathetic activation in a setting of left ventricular hypovolemia stimulates ventricular afferents that trigger hypotension and bradycardia. We tested this hypothesis by determining if excessive sympathetic activation precedes development of neurally mediated syncope, and if this correlates with alterations in baroreflex function. We studied the changes in intraarterial blood pressure (BP), heart rate (HR), central venous pressure (CVP), muscle sympathetic nerve activity (MSNA), and plasma catecholamines evoked by upright tilt in recurrent neurally mediated syncope patients (SYN, 5+/-1 episodes/mo, n = 14), age- and sex-matched controls (CON, n = 23), and in healthy subjects who consistently experienced syncope during tilt (FS+, n = 20). Baroreflex responses were evaluated from changes in HR, BP, and MSNA that were obtained after infusions of phenylephrine and sodium nitroprusside. Compared to CON, patients with SYN had blunted increases in MSNA at low tilt levels, followed by a progressive decrease and ultimately complete disappearance of MSNA with syncope. SYN patients also had attenuation of norepinephrine increases and lower baroreflex slope sensitivity, both during tilt and after pharmacologic testing. FS+ subjects had the largest decrease in CVP with tilt and had significant increases in MSNA and heart rate baroreflex slopes. These data challenge the view that excessive generalized sympathetic activation […]

Find the latest version: https://jci.me/119463/pdf

Sympathetic and Baroreceptor Reﬂex Function in Neurally Mediated Syncope Evoked by Tilt

Rogelio Mosqueda-Garcia,*‡§ Raffaello Furlan,‡ Roxana Fernandez-Violante,‡ Tushar Desai,* Marie Snell,* Zoltan Jarai,* Vasu Ananthram,§ Rose Marie Robertson,*§ and David Robertson*‡§ *Syncope Service in the Autonomic Dysfunction Unit and ‡Division of Clinical Pharmacology, §Department of Medicine, Vanderbilt University Medical Center, Nashville, Tennessee 37232-2195

Abstract bradycardia (1). In susceptible individuals, this syndrome evolves in response to relative circulating hypovolemia or dur- The pathophysiology of neurally mediated syncope is poorly ing emotional overstimulation (2). Syncope resembling that understood. It has been widely assumed that excessive sym- termed “neurally mediated syncope” can also be evoked in pathetic activation in a setting of left ventricular hypovo- otherwise healthy subjects during prolonged upright tilt (3, 4), lemia stimulates ventricular afferents that trigger hypotension infusion of vasodilator substances (5), or during pronounced and bradycardia. We tested this hypothesis by determining lower body negative pressure (6, 7). It has been assumed that if excessive sympathetic activation precedes development of patients who are prone to this type of syncope have baroreflex neurally mediated syncope, and if this correlates with alter- mechanisms and hemodynamic responses qualitatively similar ations in baroreflex function. We studied the changes in in- to those of healthy subjects and differ only quantitatively traarterial blood pressure (BP), heart rate (HR), central (8, 9). venous pressure (CVP), muscle sympathetic nerve activity Based on experimental observations in nonfainting normal (MSNA), and plasma catecholamines evoked by upright tilt volunteers, it has been proposed that an initial exaggerated in recurrent neurally mediated syncope patients (SYN, 5Ϯ1 sympathetic activation, in combination with relative ventricu- episodes/mo, n ϭ 14), age- and sex-matched controls (CON, lar hypovolemia, stimulates myocardial ventricular afferents. n ϭ 23), and in healthy subjects who consistently experi- Stimulation of these afferents may result in reflex vasodilation enced syncope during tilt (FSϩ, n ϭ 20). Baroreflex re- with or without bradycardia (10–12). Alternatively, others have sponses were evaluated from changes in HR, BP, and MSNA proposed that a sudden central resetting of baroreflexes pro- that were obtained after infusions of phenylephrine and so- duces loss of sympathetic vascular control with subsequent dium nitroprusside. Compared to CON, patients with SYN syncope (13). had blunted increases in MSNA at low tilt levels, followed Some studies have suggested that the vasodilatation seen in by a progressive decrease and ultimately complete disap- neurally mediated syncope results from a withdrawal of sym- pearance of MSNA with syncope. SYN patients also had at- pathetic tone (14–16). Consistent with this view, there is evi- tenuation of norepinephrine increases and lower baroreflex dence of reduced cardiac and renal norepinephrine (NorEpi)1 slope sensitivity, both during tilt and after pharmacologic spillover in patients who fainted during cardiac catheterization testing. FSϩ subjects had the largest decrease in CVP with (17). In addition, sympathetic withdrawal has been observed tilt and had significant increases in MSNA and heart rate using microneurography in healthy subjects undergoing infu- baroreflex slopes. These data challenge the view that exces- sion of vasodilator agents (2, 18) or during pronounced lower sive generalized sympathetic activation is the precursor of body negative pressure (6). There is only one report studying the hemodynamic abnormality underlying recurrent neu- sympathetic nerve traffic in a patient with a history of vasode- rally mediated syncope. (J. Clin. Invest. 1997. 99:2736–2744.) pressor attacks (19). In this report, disappearance of muscle Key words: neurocardiogenic syncope • vasovagal • micro- sympathetic nerve activity preceded a rather prolonged hy- neurography • tilt • baroreflex potensive attack, presumably triggered by emotional stress. To our knowledge, there are no studies in patients with Introduction recurrent neurally mediated syncope that characterize the changes in sympathetic nerve activity during the upright pos- Neurally mediated (vasovagal or neurocardiogenic) syncope ture, the position in which syncope typically occurs (2). Simi- results from excessive arteriolar dilatation and inappropriate larly, there is no information that correlates temporal changes in sympathetic nerve activity and NorEpi levels with the development of syncope or with changes in central venous pressure. Part of these results were presented in abstract form at the American There is also very little systematic information regarding po- Heart Association Annual Scientific Sessions in Atlanta, GA, 8–11 tential changes in baroreflex function in these patients. In the November, 1993 (1993. Circulation [Suppl.]. 88:I-84). present studies, in addition to evaluating sympathetic outflow Address correspondence to Dr. Rogelio Mosqueda-Garcia, Clini- during upright tilt, we studied reflex cardiovascular function in cal Research Center, MCN, Vanderbilt University Medical Center, normotensive controls, patients with a medical history of neu- Nashville, TN 37232-2195. Phone: 615-936-1750; FAX: 615-936-1906; rally mediated syncope, and healthy subjects who experienced E-mail: [email protected] syncope only during tilt. Received for publication 12 September 1996 and accepted in re- vised form 13 March 1997.

J. Clin. Invest. 1. Abbreviations used in this paper: CVP, central venous pressure; © The American Society for Clinical Investigation, Inc. Epi, epinephrine; FSϩ, subjects with false-positive tilt; HR, heart 0021-9738/97/06/2736/09 $2.00 rate; MSNA, muscle sympathetic nerve activity; NorEpi, norepineph- Volume 99, Number 11, June 1997, 2736–2744 rine; SYN, patients with recurrent syncope.

2736 Mosqueda-Garcia et al.

Methods fossa. In the opposite arm, an intravenous line was positioned for blood sampling or drug administration. Study groups and patient selection. The subjects abstained from caf- Muscle sympathetic nerve activity (MSNA) was recorded from feine-containing products and smoking for at least 7 d. All medica- the right peroneal nerve using the microneurography technique (20, tions were stopped for at least five drug half-lives before the study. 21) described elsewhere (22, 23). In brief, the right leg was immobi- The first group of patients included 14 subjects (5 males and 9 fe- lized and securely positioned on the tilt table using an adjustable leg males, age range ϭ 21–41 yr, mean ϭ 34Ϯ1.32 yr) suffering from re- brace to maximize successful recordings. A unipolar tungsten elec- current syncope (SYN, two or more episodes per month) who were trode (uninsulated tip diameter of 1–5 ␮m, shaft diameter of 200 ␮m) recruited from the Vanderbilt Autonomic Dysfunction Center. In this was inserted into the muscle nerve fascicles of the peroneal nerve for group, the diagnosis was established after a review of the patient’s multiunit recordings. Nerve activity was fed to a preamplifier (1,000- medical history, a comprehensive physical examination, and a tilt ta- fold amplification) and filtered using a bandwidth between 700 and -fold, and in-100 ف ble study that indicated the presence of this syndrome. In addition, 2,000 Hz. The signal was then rectified, amplified these patients had negative cardiologic (12-lead EKG, echocardio- tegrated (time constant ϭ 0.1 s) to obtain a mean voltage display of gram, and ambulatory electrocardiographic monitoring), neurologic sympathetic activity. We considered a recording of sympathetic nerve (electroencephalogram and computed axial tomography assessment), activity acceptable when (a) the microneurography signal demon- endocrinologic (fasting blood glucose and thyroid function), hemato- strated spontaneous, pulse-coupled bursts of nerve activity that ex- logic, and psychiatric evaluations. The second group (FSϩ) was com- ceeded by threefold the background noise signal; (b) when there was prised of 20 subjects (11 males and 9 females, 33Ϯ1.86 yr) who had no no evidence of increased activity over time; and (c) when activity re- history of syncope and had an unremarkable medical history and turned to basal after procedures affecting sympathetic tone. The sym- physical examination, but nevertheless consistently experienced pathetic nature of the recordings was indicated by increased activity fainting during tilt. The subjects were identified during the recruit- during the hypotensive phase of the Valsalva maneuver or by inhibi- ment of normal control subjects for tilt studies. The final group con- tion of the activity during the blood pressure overshoot of the same sisted of healthy normotensive controls (CON, n ϭ 23) who had no maneuver. From the 26 patients undergoing invasive tilt, MSNA re- syncope or presyncope on tilt. This group included at least one sub- cordings were incomplete, lost, or inadequate in 12 subjects. These ject who was age- and sex-matched for each patient. are not included in the final analysis. In the remaining 14, the record- Experimental protocol. The experimental protocol was approved ing electrode did not dislodge, and intact tracings were obtained dur- by the Vanderbilt University Institutional Review Board in Human ing the entire tilt, including pre- and postsyncope periods. In five of Research, and all the subjects gave written informed consent before the completed syncope patients, the recording electrode was subse- any procedure was initiated. An initial 75Њ tilt table study was per- quently removed to allow leg mobility. The electrode was reposi- formed (Fig. 1) in which blood pressure was monitored with a Fina- tioned later on for performance of the baroreflex study. Data were press device (Ohmeda, Austell, GA). Patients were excluded if this stored using a digitizing system, and were stored on VCR tape for tilt study was normal. If the study was positive (development of syn- subsequent analysis. cope or presyncope resulting from hypotension, bradycardia, or both), Tilt procedure. Basal recordings were made 30 min after instru- the patients were reevaluated 1 wk later with another tilt study. If mentation, including a blood sample for catecholamine determina- the response was similar to the first tilt, the patients were included in the tion. Subsequently, the subjects underwent an upright tilt at 15Њ inter- remaining portion of the study. A similar approach was used for the in- vals every 3 min until reaching 75Њ. The controls were maintained in clusion of the control subjects, except that in these individuals, the this position until 30 min of tilt were completed; the patients and FSϩ two tilt tests were negative. The FSϩ group was formed by subjects were placed in supine positions when presyncope or syncope devel- with normal medical histories and physical findings, but with consis- oped. Samples for plasma catecholamines were obtained at the end of tent positive tilt studies. each tilt interval and every 5 min at the maximal tilt angle or, in the For the last part of the study, subjects were admitted to the Gen- corresponding subjects, immediately after syncope or presyncope. eral Clinical Research Center. After overnight fasting, they under- Baroreflex testing. After the tilt, subjects were allowed to rest for went another tilt. Blood pressure (BP) was recorded from an intra- the next 45 min and then, while the subjects were supine, another arterial line placed in the radial artery, and heart rate (HR) was baseline for hemodynamics and MSNA was recorded for a 3-min pe- obtained from lead II of the surface electrocardiogram. Central venous riod. After this, the subjects received graded infusions of phenyleph- pressure (CVP) was obtained using a high fidelity 3-French Millar rine (0.1–3.2 ␮g/kg per min) or sodium nitroprusside (0.1–2.8 ␮g/kg Mikro-Tip transducer (Millar Instruments, Inc., Houston, TX) placed per min) with the order of the two agents randomized. Infusions were through the median or basilic vein at the level of the antecubital given in ascending order, with each concentration lasting 3 min; the

Figure 1. Flowchart of the experimental protocol. A plus sign indicates medical findings compatible with neurally mediated syncope or the development of presyncope or syncope during the tilt table test. A minus sign indicates negative medical findings or a normal tilt table test. An asterisk indicates loss, incomplete, or inadequate recording of muscle sympathetic nerve activity during the tilt. Invasive tilt represents the test done with intraarterial recording of blood pressure, microneurography, and CVP determina- tion. n, the number of subjects undergoing each procedure or the number excluded after it.

Sympathetic and Baroflex Function in Syncope 2737

concentration was increased until BP rose with phenylephrine by at least 40/30 mmHg (systolic/diastolic, respectively) or fell with sodium nitroprusside by at least 30/20 mmHg. Definitions. We defined neurally mediated syncope as a transient loss of consciousness that was not compatible with other altered states of consciousness in the patient’s history. Presyncope was defined as a prodromal period characterized by signs and symptoms of imminent syncope. These included severe lightheadedness preceded by nausea or yawning, severe weakness, transient graying or “tunnel” vision, and/or hearing loss. A positive response to tilt was defined as the development of syncope or presyncope caused by hypotension, bradycardia, or both. Analyses. For the evaluation of the hemodynamic responses during tilt, we analyzed the entire last minute of each tilt angle and four different 1-min periods during the 75Њ position. Measurements of MSNA were made from the tracings of the microneurogram using a digitizer tablet, which uses computer software for data storage and analysis. Results were evaluated as the number of bursts per minute or as total MSNA (arbitrary units) calculated as bursts per minute times the mean burst amplitude. Since no significant qualitative differences were seen between these two methods of analysis, results are presented as percentage changes in MSNA (bursts per minute) from baseline values. Concentrations of plasma NorEpi and epinephrine (Epi) were obtained using an HPLC method reported previously (24). Quantification of baroreflex responses to pharmacological ma- nipulations of blood pressure used least squares regression analysis, as reported previously (25). Intervals between each R wave of the EKG (R-R intervals) and sympathetic nerve activity were averaged during the maximal change in BP. The sigmoid relationship between incremental averages of systolic BP and R-R intervals were then plotted, and the linear portion of these curves was used to derive a slope and a correlation coefficient. A similar process was applied to the relationship between incremental averages of diastolic BP and MSNA. Figure 2. Hemodynamic and microneurographic responses to tilt in All data are expressed as meanϮSEM. Comparisons were ana- the control group. The diagram at the top presents the changes in sys- lyzed by Student’s t test for unpaired observations or ANOVA fol- tolic BP (SBP, filled circles), diastolic BP (DBP, filled triangles), HR lowed by Dunnett’s test for significant differences, as appropriate. (filled squares), and CVP (filled diamonds) during different levels of P Ͻ 0.05 was considered significant. tilt. The diagram at the bottom presents the changes in sympathetic nerve traffic (⌬MSNA) recorded as percentage change over baseline Results in bursts/min. For both diagrams, there is one value for 15Њ, one for 30Њ, two for 45Њ, three for 60Њ, and four for 75Њ. The vertical lines rep- Clinical characteristics. The characteristics of the study groups resent SEM. are presented in Table I. In the patients with recurrent syncope, the duration of symptoms ranged from 7 mo to 10 yr, with a frequency of 5Ϯ1.05 episodes per month. None of the controls or FSϩ subjects had contributory medical history or Hemodynamic and microneurographic responses to tilt. physical findings. Age, weight, diastolic BP, and HR were not The hemodynamic responses to tilt in controls were character- significantly different among the groups. Intraarterial systolic ized by a minimal decrease in systolic BP (maximal change at BP was lower in subjects with recurrent syncope when com- the end of tilt was Ϫ9.68Ϯ3.04 mmHg), a minor increase in di- pared to the false positive and control groups. Supine plasma astolic BP (4.2Ϯ1.7 mmHg), and an initial angle-related de- NorEpi and Epi values were not different among the study crease in CVP that remained stable after reaching 75Њ (maxi- groups (Table I). mal decrease of Ϫ7.5Ϯ1.8 mmHg; Fig. 2). HR progressively

Table I. Clinical Characteristics of the Study Subjects

Age Sex Weight SBP DBP HR MSNA NorEpi Epi n

yr kg mmHg mmHg bpm bursts/min pg/ml pg/ml

CON 33Ϯ1.59 8M/15F 75.87Ϯ1.94 144Ϯ3.68 69Ϯ1.78 65Ϯ1.94 16Ϯ1.6 237Ϯ18 22Ϯ523 SYN 34Ϯ1.32 5M/9F 70.71Ϯ2.89 123Ϯ4.34* 66Ϯ1.37 68Ϯ3.54 13Ϯ2.73 206Ϯ20 14Ϯ314 FSϩ 33Ϯ1.86 11M/9F 69.10Ϯ2.46 139Ϯ4.19 65Ϯ2.19 64Ϯ2.57 16Ϯ1.85 222Ϯ27 60Ϯ34 20

CON, control group; SYN, group with recurrent syncope; FSϩ, subjects with false-positive tilt; SBP, systolic blood pressure; DBP, diastolic blood pressure. *Significantly different from CON (P Ͻ 0.001) and FSϩ (P Ͻ 0.007) groups.

2738 Mosqueda-Garcia et al.

Figure 3. Supine and maximal upright values of plasma NorEpi and Epi during the tilt study. Plasma samples for supine (SUP) NorEpi and Epi were collected in prechilled tubes after the subjects rested at least 30 min after instrumentation. The maximal values obtained for both catecholamines during tilt (TLT) are presented in this figure. The open symbols represent the individual values for all the partici- pants in the study. The filled symbols at the left and right of the open Figure 4. Hemodynamic and microneurographic responses to tilt in symbols indicate the meanϮSEM of each group that was control patients with recurrent syncope. Arrangement, abbreviations, and (CON), false positive (FSϩ), and patients with recurrent syncope symbols are similar to those in Fig. 2. For both diagrams, there is one (SYN). Note that the maximal NorEpi values were significant lower value for 15Њ, one for 30Њ, two for 45Њ, three for 60Њ, and two for 75Њ. in the SYN group when compared to CON (P Ͻ 0.0003) and in the Note that when compared to Fig. 2, there is a different scale for the FSϩ group when compared to CON (P Ͻ 0.002). y axis in the upper tracing and the absence of almost any increase in MSNA during the initial phases of tilt. increased until reaching a maximal change of ϩ32Ϯ2.4 bpm at the end of the tilt. Muscle sympathetic nerve activity in controls increased in addition, the maximal increase in sympathetic activity (which response to tilt and, in general, correlated with the degree of was clearly blunted) was seen earlier, and a phase then fol- tilt angle. The increase in MSNA was maximal within 10 min at lowed during which a progressive decrease in MSNA resulted 75Њ (277Ϯ87% burst /min, Fig. 2) and remained elevated, al- in total neural silence immediately preceding syncope (Figs. 4 though lower than maximum, during the rest of the study. and 5). Similar differences were observed when the levels of Plasma venous NorEpi and Epi followed a pattern similar to plasma venous NorEpi were compared between controls and the MSNA with a maximal increase more than threefold over patients. Although supine basal levels were similar, the maxi- basal values (Fig. 3). mal increase in plasma venous NorEpi in patients (378Ϯ56 pg/ Patients with recurrent syncope developed significant hy- ml) was significantly less than in controls (637Ϯ52 pg/ml, P Ͻ potension, even at low levels of tilt (Fig. 4), which progres- 0.0003). On the other hand, in the recurrent syncope group, sively worsened and resulted in syncope between 60Њ and 75Њ. plasma Epi levels were significantly higher than in controls The decrease in BP during tilt in these patients was associated (Fig. 3). with some tachycardia and decreases in CVP that were not The tilt study in FSϩ subjects was characterized by a signif- significantly different from values observed in controls (com- icant decrease in BP (evident from 30Њ), which resulted in pro- pare Figs. 2 and 4). For instance, during the first minute of 45Њ nounced hypotension and syncope between 60Њ and 75Њ tilt. In tilt, CVP and systolic and diastolic BP had decreased by these subjects, the increase in HR reached a maximal change Ϫ3.2Ϯ0.6, Ϫ15.2Ϯ3.4, and Ϫ6.2Ϯ3.4 mmHg, respectively, with of 25.3Ϯ3.7 bpm, and the maximal decrease in CVP was larger a minimal heart rate increase of 10Ϯ3 bpm (Fig. 4). than that recorded in the other two groups (Fig. 6). When More pronounced differences between controls and pa- compared to controls, the changes in MSNA were character- tients with recurrent syncope were observed when the sympa- ized by similar increments during the low levels of tilt, as well thetic responses to tilt were analyzed. In these patients, the in- as a tendency for a more pronounced increment at the 60Њ and crease in MSNA was severely impaired at low levels of tilt. In early 75Њ tilt. These changes were followed by an abrupt disap-

Sympathetic and Baroflex Function in Syncope 2739 Figure 5. Integrals of microneurographic recordings while supine and during tilt of a control subject, a patient with recurrent syncope (REC. SYNCOPE), and one subject with consistent syncope evoked by tilt (FALSE ϩ). Recordings were obtained from the right peroneal nerve without moving the recording electrode during the tilt study. Note that in the control subject, the significant increase in MSNA at 15Њ and 30Њ was followed by a less pronounced increment between 45Њ and 75Њ. In the recurrent syncope patient, there was almost no increase in MSNA at low levels of tilt, and the maximal increment was observed at the first minute of 60Њ. This was followed by a progressive decrease in MSNA until total disappearance and syncope. The artifacts (asterisk) recorded in this patient were produced by involuntary contraction of leg muscles during syncope. In spite of this, note that MSNA reappeared during the recovery period. In the false-positive subject, the maximal increase in MSNA was followed by a more abrupt decrease with total withdrawal and syncope. As with in the middle tracing, the asterisk indicates recording artifacts. The arrows represent the approxi- mate time of syncope. pearance in MSNA that preceded the development of syn- served in subjects with recurrent syncope, while the increase in cope. Plasma NorEpi and Epi also increased in this group of plasma Epi was more than in controls (P Ͻ 0.025). subjects (Fig. 3). The absolute maximal increase in NorEpi was Reflex cardiovascular responses. Correlation analysis of the less than in controls (P Ͻ 0.002) and similar to the one ob- hemodynamic changes evoked by tilt (R-R interval versus systolic BP; MSNA versus CVP) indicated significant differences among the three study groups. In controls, the changes between R-R interval and systolic blood pressure yielded a slope of 31.2 ms/mmHg (r ϭ 0.59; Fig. 7). In the false-positive group, the slope was 43.7 ms/mmHg (r ϭ 0.76), while in the group of

Figure 7. Relationship between changes in R-R interval and systolic BP during upright tilt. The figure presents the plotted values and re- Figure 6. Hemodynamic and microneurographic responses to tilt in gression lines obtained from correlating the changes in R-R interval false-positive subjects. Arrangement, abbreviations, and symbols are (⌬ R-R) with the changes in systolic BP (⌬ SBP) in the three study similar to those in Fig. 2. For both diagrams, there is one value for groups. The filled circles, hollow circles, and filled triangles represent 15Њ, one for 30Њ, two for 45Њ, three for 60Њ, and three for 75Њ. Note that the values obtained in the control subjects (CON), false-positive sub- when compared to Figs. 2 and 3, there is a different scale for the y axis jects (FSϩ), and recurrent syncope patients (SYN), respectively. Val- in the upper tracing, while the scale for MSNA is the same. ues between parentheses indicate the regression slope.

2740 Mosqueda-Garcia et al. Figure 8. Relationship between changes in muscle sympathetic nerve activity and changes in CVP during upright tilt. The symbols represent the plotted values and regression lines obtained from correlating Figure 9. Baroreflex slopes obtained in the study subjects from infu- the changes in MSNA (⌬ MSNA) with changes in CVP (⌬ CVP). sions of phenylephrine and sodium nitroprusside. The bars represent Symbols are the same as in Fig. 7. the mean slope value from correlating the changes in R-R interval with systolic BP (slope ms/mmHg) or the changes in MSNA with diastolic BP (slope burst/min per mmHg) evoked by ascending infusions of phenylephrine and sodium nitroprusside. The open bars represent the slope values in controls (CON), while the right ascending patients with recurrent syncope, the slope of the curve was sig- hatched and cross-hatched bars represent the slope values for pa- nificantly reduced (13.2 ms/mmHg; r ϭ 0.84; Fig. 7). The cor- tients with recurrent syncope (SYN) and false-positive subjects relation between CVP and MSNA resulted in slope values of (FSϩ), respectively. The vertical lines in the middle of each column Ϫ34.4 bursts/min per mmHg for controls (r ϭ 0.76), Ϫ19.5 represent SEM change from baseline. The horizontal lines indicate bursts/min per mmHg for the false positive group (r ϭ 0.90), the difference between the corresponding groups with single and and Ϫ5.71 bursts/min per mmHg for patients with recurrent double asterisks representing P Ͻ 0.05 and P Ͻ 0.01, respectively. syncope (r ϭ 0.15; Fig. 8). Using stepwise infusions of phenylephrine and sodium nitroprusside, we further documented alterations in the reflex control of the circulation in subjects with syncope. Among the three groups, patients with recurrent syncope had the lowest tions in circulating volume (as indicated from the decrease in baroreflex slopes for both correlations, R-R interval versus CVP values), with consequent falls in cardiac output and inap- systolic BP, and MSNA versus diastolic BP (Fig. 9). No signifi- propriate reflex tachycardia. In addition, these patients also cant differences were observed between controls and false- exhibited reduced maximal increases in MSNA and plasma positive subjects for the MSNA versus diastolic BP slopes NorEpi that were inadequate to compensate for the significant (Fig. 9). fall in arterial BP. This was then followed by progressive MSNA inhibition until complete disappearance and syncope. Discussion Attempts to evaluate sympathetic function with plasma NorEpi in neurally mediated syncope have produced variable and This study reports two major observations relevant to the seemingly contradictory results; NorEpi concentrations have pathophysiology of neurally mediated syncope. First, there are been reported to be either relatively low (13, 26, 27), normal important and well-defined differences between patients with (28), or moderately elevated (4, 29, 30). Confounding factors recurrent syncope and healthy normal subjects who faint only such as time of sampling, potential changes in NorEpi clear- when tilted. Second, significant alterations in the reflex control ance, and alterations in spillover to the general circulation are of the circulation in recurrent syncope patients contribute to likely to account for these contradictory results (17, 31). More the orthostatic intolerance experienced by these subjects. recently, Jacobs and collaborators have documented a blunted Our results indicate that the initial sympathetic response to noradrenergic response combined with exaggerated adreno- changes in circulating volume is different, not only between medullary activation in subjects who subsequently developed controls and patients, but also among different individuals vasodepressor reactions during exposure to pronounced lower experiencing syncope. In the group with recurrent syncope, body negative pressure (13). the initial increase in MSNA during early tilt was severely The initial MSNA response pattern in patients with recur- reduced. Hemodynamically, these patients had significant de- rent syncope recorded in this study seems to be at variance creases in BP that probably resulted from orthostatic reduc- with earlier published MSNA recordings. It is noteworthy,

Sympathetic and Baroflex Function in Syncope 2741 however, that except for one (19), all the microneurographic clines during continuous electrical stimulation of the carotid si- recordings in previous reports were obtained from subjects nus nerve (46; equivalent to baroreceptor loading) in humans, without clinical evidence or history of syncope (6, 18, 32). In there has not been clinical experimental evidence demonstrat- the study documenting microneurographic changes in a sub- ing activation of high pressure baroreceptors at very low arte- ject with vasodepressive attacks, the patient had extremely rial pressures. long periods of hypotension (up to 2 h) and was studied only Implications and limitations of the study. Our findings indi- while supine. Although this case may represent a form of “cen- cate that in a significant proportion of patients with recurrent tral” neurally mediated syncope, in a majority of patients, syn- neurally mediated syncope, alterations of neural-cardiovascu- cope occurs with upright posture and is of short duration (2). lar reflexes result in an inability to adequately respond to On the other hand, the response pattern in false-positive sub- orthostatic stress. It may be argued, on the other hand, that a jects supports the view that pronounced MSNA increases oc- single mechanism (activation of cardiac inhibitory afferents) is cur before sudden withdrawal. Together, these observations responsible for the final sympathetic withdrawal and that the indicate a heterogeneity in the initial sympathetic response to differential changes in sympathetic activity represent only dif- orthostatic stress in subjects prone to fainting with tilt (but oth- ferent ends of the spectrum of “normal” response (i.e., sub- erwise healthy) and in patients with recurrent syncope. Addi- jects in the SYN group had increased sensitivity to sympathetic tional evidence for differential sympathetic responses in neu- stimulation of their cardiac afferents while others required rally mediated syncope is provided by studies assessing pronounced increases). This explanation, however, does not spectral analysis of HR variability. These studies have indi- account for the altered baroreflex function and the absence of cated increased (33, 34), reduced (35–37), or even nonsus- initial sympathetic activation in response to tilt in several of tained noradrenergic activation (38) preceding syncope. the present cases. Furthermore, there is emerging evidence ar- Alterations in baroreflex function have been implicit, al- guing against the ventricular theory as the universal mecha- though not always discussed, when examining the final cardio- nism underlying neurally mediated syncope. For instance, us- vascular response of neurally mediated syncope. The concomi- ing echocardiographic measurements of left cardiac chamber tant presence of hypotension and inadequate HR response size and determinations of stroke volume, Novak et al. (47) indicate that the normal baroreflex function has been super- found no evidence of progressive excessive cardiac emptying seded. Some authors have proposed that “sudden” resetting before the onset of syncope in susceptible patients. More re- baroreflexes results in sympathoneural withdrawal despite sig- cently, Lui et al. (48) assessed left ventricular wall stress and nificant hypotension (13). Our results now clearly document segmental wall thickening preceding syncope and found no ev- that an underlying baroreceptor reflex dysfunction leads to a idence compatible with activation of left ventricular mecha- defective sympathetic response to orthostatic stress in patients noreceptors. with recurrent syncope. In these patients, significant reduc- Assessment of the pathophysiological mechanisms leading tions in reflex responses were observed, not only during tilt to neurally mediated syncope has been difficult, since the (changes in CVP versus MSNA and R-R versus systolic BP), spontaneous expressions of this syndrome occur somewhat but also when baroreflex activity was evaluated with phenyl- randomly and usually have resolved by the time that medical ephrine and sodium nitroprusside infusions. The decreased evaluation is available. Although the use of tilt table testing baroreflex function explains the inability of these subjects to has facilitated the study of the hemodynamic characteristics in increase sympathetic outflow in response to reductions in pres- this type of syncope, questions remain regarding the specific- sure. Recent reports by at least two different groups of investi- ity, sensitivity, and variability of this test (49). In our protocol, gators (39, 40) have documented similar impairment of barore- we had strict selection criteria that allowed the study of sub- ceptor reflex sensitivity in different populations of subjects jects with consistent responses to upright tilt. In their respec- with neurally mediated syncope. In one of these studies, the tive groups, subjects had either positive or negative tilt table impairment of baroreflex function predicted the development tests. Individuals in whom the response to tilt was variable of syncope during tilt (40). Furthermore, we have recently (positive/negative or negative/positive) were not studied. Al- been able to manipulate the efferent sympathetic limb of the though a remote possibility, this selection process may have baroreflex in some of these SYN patients (41) and confirm excluded subjects with syncope in whom different pathophysi- that an enhancement in overall sympathetic tone reduces ological mechanisms are present. Similarly, our observations orthostatic intolerance and prevents syncope induced by tilt. relate to posture-induced syncope and might have no rele- Taken together, these findings support the concept that im- vance to other types of neurally mediated syncope (i.e., the pairment of baroreflex function is a mechanism that leads to central type, such as emotional fainting). abnormal sympathetic responses to standing in susceptible in- Finally, the presence of either reduced or increased MSNA dividuals. in subjects with syncope may also explain the dissimilar thera- More difficult to clarify is the progressive decrease in MSNA peutic response observed in different populations (50, 51). during tilt. One possibility is that in susceptible subjects, the Conceivably, patients in whom syncope is associated with a failure to sense changes in pressure may be further compli- blunted increase in sympathetic tone would benefit from ␣-adren- cated by resetting of baroreflexes, which results in sympathetic ergic receptor agonists or from agents that increase sympa- withdrawal. In laboratory animals, resetting leading to sympathetic tone. In contrast, subjects with exaggerated noradrener- thetic inhibition has been described during severe hemorrhage gic changes would respond to ␤ blockers, while those with (42) and hypoglycemia (43). In humans, plasma NorEpi first pronounced orthostatic decreases in circulatory volume might increases and then decreases during progressive reductions of benefit from mineralocorticoids. Indeed, we believe that our arterial pressure (44). This phenomenon has been ascribed to results provide the basis for a more rational design of thera- paradoxical activation of baroreceptors during pronounced peutic trials in neurally mediated syncope. Our patients with hypotension (45). While the degree of inhibition of MSNA de- recurrent syncope have responded positively to agents that in-

2742 Mosqueda-Garcia et al. crease sympathetic tone, such as direct ␣-adrenergic vascular 19. Yatomi, A., A. Iguchi, K. Uemura, N. Sakamoto, S. Iwase, and T. Mano. agonists (midodrine), or to predominantly centrally acting agents 1989. A rare case of recurrent vasodepressive attacks of 2-hour duration: analysis of the mechanism by muscle sympathetic nerve activity recording. Clin. Car- (i.e., yohimbine), which increase sympathetic tone, whereas diol. 12:164–168. they failed treatment with ␤-adrenergic blockers (propranolol, 20. Vallbo, A.B., K.E. Hagbarth, H.E. Torebjork, and B.G. Wallin. 1979. atenolol, etc.). It is therefore reasonable to suggest that the de- Somatosensory, proprioceptive and sympathetic activity in human peripheral nerves. Physiol. Rev. 59:919–957. lineation of pathophysiological mechanisms leading to syn- 21. Sundlof, G., and B.G. Wallin. 1978. Human muscle nerve sympathetic cope in individual patients will result in better and more ratio- activity at rest. Relationship to blood pressure and age. J. Physiol. 274:621–637. nal treatment. This must also be considered when evaluating 22. Mosqueda-Garcia, R. 1996. Microneurography in neurological research. Am. Acad. Neurol. (Autonom. Nerv. Sys. Sec.). 2:4–5. the efficacy of different pharmacological agents in compara- 23. Davis, S.N., C. Shavers, F. Costa, and R. Mosqueda-Garcia. 1996. Role tive trials. of cortisol in the pathogenesis of deficient counterregulation after antecedent In summary, we have found that the sympathetic response hypoglycemia in normal humans. J. Clin. Invest. 98:680–691. 24. Rea, R., I. Biaggioni, R.M. Robertson, V. Haile, and D. Robertson. in subjects susceptible to neurally mediated syncope is hetero- 1990. Reflex control of sympathetic nerve activity in dopamine ␤-hydroxylase geneous. The finding of altered baroreceptor reflex function in deficiency. Hypertension (Dallas). 15:107–112. a subpopulation of syncope patients suggests the potential par- 25. Mosqueda-Garcia, R., C. Tseng, I. Biaggioni, and D. Robertson. 1990. Effects of caffeine on baroreflex activity in man. J. Pharmacol. Exp. Ther. 255: ticipation of multiple peripheral and/or central mechanisms 374–381. that eventually lead to severe hypotension and loss of con- 26. Goldstein, D.S., M. Spanarkel, A. Pitterman, R. Toltzis, E. Gratz, S. Ep- sciousness. stein, and H.R. Keiser. 1982. Circulatory control mechanisms in vasodepressor syncope. Am. Heart J. 104:1071–1075. 27. Ziegler, M.G., C. Echon, K.D. Wilner, P. Specho, C.R. Lake, and J.A. Acknowledgments McCutchen. 1986. Sympathetic nervous withdrawal in the vasodepressor (vasovagal) reaction. J. Autonom. Nerv. Syst. 17:273–278. 28. Sra, J.S., V. Murthy, A. Natale, M.R. Jazayeri, A. Dhala, S. Deshpande, We deeply appreciate the editorial assistance of Mrs. Dorothea Boe- M. Sheth, and M. Akhtar. 1994. Circulatory and catecholamine changes during mer in the preparation of this document. head-up tilt testing in neurocardiogenic (vasovagal) syncope. Am. J. Cardiol. This work was supported in part by a National Institutes of 73:33–37. Health (NIH) Program Project Grant grant HL46681, a grant from 29. Chosy, J.J., and D.T. Graham. 1965. Catecholamines in vasovagal fainting. J. Psychosom. Res. 9:189–194. the Smokeless Tobacco Research Council, and the Vanderbilt Uni- 30. Fitzpatrick, A., T. Williams, R. Ahmed, S. Lightman, S.R. Bloom, and versity General Clinical Research Center NIH grant RR00095. R. Sutton. 1992. Echocardiographic and endocrine changes during vasovagal syncope induced by prolonged head-up tilt. Eur. J. Cardiac Pacing Electrophys. 2:121–128. References 31. Goldstein, D.S. Catecholamines in plasma and cerebrospinal fluid: sources and meanings. In Brain Peptides and Catecholamines in Cardiovascular 1. Glick, G., and P.N. Yu. 1963. Hemodynamic changes during spontaneous Regulation. J.P. Buckley and C.M. Ferrario, editors. 15–25. vasovagal reactions. Am. J. Med. 34:42–51. 32. Scherrer, U., S. Vissing, B.J. Morgan, P. Hanson, and R.G. Victor. 1990. 2. Van Lieshoutt, J.J., W. Wieling, J.M. Karemaker, and D.L. Eckberg. Vasovagal syncope after infusion of a vasodilator in a heart-transplant patient. 1991. The vasovagal response. Clin. Sci. 81:575–586. N. Engl. J. Med. 322:602–604. 3. Fitzpatrick, A., G. Theodorakis, P. Vardas, R.A. Kenny, C.M. Travill, A. 33. Pruvot, E., J.M. Vesin, J. Schlaepfer, M. Fromer, and L. Kappenberg. Ingram, and R. Sutton. 1991. The incidence of malignant vasovagal syndrome in 1994. Autonomic imbalance assessed by heart rate variability analysis in vaso- patients with recurrent syncope. Eur. Heart J. 12:389–394. vagal syncope. Pacing & Clin. Electrophysiol. 17:2201–2206. 4. Sander-Jensen, K., N.H. Secher, and A. Astrup. 1986. Hypotension in- 34. Yamamoto, M., H. Sato, S. Abe, T. Namiki, H. Iinuma, L. Fu, T. duced by passive head-up tilt: endocrine and circulatory mechanisms. Am. J. Aizawa, and K. Kato. 1993. Unusual increase in sympathetic nervous activity Physiol. 251:R742–R748. preceding syncope in patients with neurally mediated syncope. Circulation. 5. Robertson, D., G.A. Johnson, R.M. Robertson, A.S. Nies, D.G. Shand, 88(Suppl. I):I-44. and J.A. Oates. 1979. Comparative assessment of stimuli that release neuronal 35. Baharav, A., M. Mimouni, T. Lehrman-Sagie, S. Izraeli, and S. Aksel- and adrenomedullary catecholamines in man. Circulation. 59:637–647. rod. 1993. Spectral analysis of heart rate in vasovagal syncope: the autonomic 6. Sander, J.S., and D.W. Ferguson. 1989. Profound sympathoinhibition nervous system in vasovagal syncope. Clin. Autonom. Res. 3:261–269. complicating hypovolemia in humans. Ann. Int. Med. 111:439–441. 36. Morillo, C.A., G.J. Klein, D.L. Jones, and R. Yee. 1994. Time and fre- 7. Murray, R.H., L.J. Thompson, J.A. Bowers, and C.D. Albright. 1968. He- quency domain analyses of heart rate variability during orthostatic stress in pa- modynamic effects of graded hypovolemia and vasodepressor syncope induced tients with neurally mediated syncope. Am. J. Cardiol. 74:1258–1262. by lower body negative pressure. Am. Heart J. 76:799–811. 37. Vardas, P.E., G.E. Kochiadakis, A.E. Orfanakis, S. Chrysostomakis, 8. Blomqvist, C.G, and H.L. Stone. Cardiovascular adjustments to gravita- and E.G. Manios. 1995. Study of the autonomic nervous system in patients with tional stress. In Handbook of Physiology. J.T. Shepherd and F.M. Abboud, edi- neurocardiogenic (vasovagal) syncope. PACE. 18:845. tors. American Physiological Society, Washington, D.C. 1025–1063. 38. Lepicovska, V., P. Novak, and R. Nadeau. 1992. Time-frequency dy- 9. Ludbrook, J. 1989. Faint heart. J. Autonom. Nerv. Sys. 60:83–92. namics in neurally mediated syncope. Clin. Autonom. Res. 2:317–326. 10. Mark, A.L. 1983. The Bezold-Jarisch reflex revisited: clinical implica- 39. Haruhiko, I., I. Shigeki, M. Mitsuru, U. Tetsuro, M. Naoki, and K. Hi- tions of inhibitory reflexes originating in the heart. J. Am. Coll. Cardiol. 1:90– roshi. 1996. Abnormal baroreflex sensitivity in patients with neurally mediated 102. syncope. Circulation. 94(Suppl. I):I-621. 11. Abboud, F.M. 1989. Ventricular syncope. Is the heart a sensory organ? 40. Morillo, C.A., K.A. Ellenbogen, M.A. Wood, L.A. Beightol, and D.L. New Engl. J. Med. 320:346–351. Eckberg. 1996. Impaired cardio-vagal and muscle sympathetic baroreflex out- 12. Waxman, M.B., D.A. Cameron, and R.W. Wald. 1993. Role of ventricu- flow in patients with head-up tilt induced neurally mediated syncope. Circula- lar vagal afferents in the vasovagal reaction. J. Am. Coll. Cardiol. 21:1138–1141. tion. 94(Suppl. I):I-544. 13. Jacobs, M.C., D.S. Goldstein, J.J. Willemsen, P. Smits, T. Thien, R.A. 41. Mosqueda-Garcia, R., R. Furlan, R. Fernandez-Violante, M. Snell, and Dionne, and J.W.M. Lenders. 1995. Neurohumoral antecedents of vasodepres- D. Robertson. 1996. Enhancement of central noradrenergic outflow prevents sor reactions. Eur. J. Clin. Invest. 25:754–761. neurally mediated syncope. Clin. Autonom. Res. 6:290. 14. Barcroft, H., O.G. Edlhom, J. McMichael, and E.P. Sharpey-Schafer. 42. Victor, R.G., P. Thoren, D.A. Morgan, and A.L. Mark. 1989. Differen- 1944. Posthaemorrhagic fainting. Study by cardiac output and forearm flow. tial control of adrenal and renal sympathetic nerve activity during hemorrhagic Lancet. i:489–491. hypotension in rats. Circ. Res. 64:686–694. 15. Weiss, S., R.W. Wilkins, and F.W. Haynes. 1937. The nature of circula- 43. Nijima, A. 1975. The effect of 2-deoxy-D-glucose and D-glucose on ef- tory collapse induced by sodium nitrite. J. Clin. Invest. 16:73–84. ferent discharge rate of sympathetic nerves. J. Physiol. 251:231–243. 16. Greenfield, A.D.M. 1966. Survey of the evidence for active neurogenic 44. Eckberg, D.L., S.W. Harkins, J.M. Fritsch, G.E. Musgrave, and D.F. vasodilation in man. Fed. Proc. Fed. Am. Soc. Exp. Biol. 25:1607–1610. Gardner. 1986. Baroreflex control of plasma norepinephrine and heart period 17. Esler, M., G. Jennings, G. Lambert, I. Meredith, M. Horne, and G. in healthy subjects and diabetic patients. J. Clin. Invest. 78:366–374. Eisenhofer. 1990. Overflow of catecholamine neurotransmitters to the circula- 45. Eckberg, D.L., and P. Sleight. 1992. Afferent baroreceptor activity. In tion: source, fate and functions. Physiol. Rev. 70:963–985. Human Baroreflexes in Health and Disease. D.L. Eckberg and P. Sleight, edi- 18. Wallin, B.G., and G. Sundolf. 1982. Sympathetic outflow to muscles dur- tors. Oxford University Press, New York. 123–152. ing vasovagal syncope. J. Autonom. Nerv. Sys. 6:287–291. 46. Wallin, B.G., G. Sundlof, and W. Delius. 1975. The effect of carotid si-

Sympathetic and Baroflex Function in Syncope 2743 nus nerve stimulation on muscle and skin nerve sympathetic activity in man. evaluating syncope: a comprehensive literature review. Am. J. Med. 97:78–88. Pflugers Archiv. Eur. J. Physiol. 358:101–110. 50. Kaufmann, H. 1995. Neurally mediated syncope: pathogenesis, diagno- 47. Novak, V., G. Honos, and R. Schondorf. 1996. Is the heart empty at syn- sis, and treatment. Neurology. 45(Suppl. 5):s12–s18. cope? J. Autonom. Nerv. Sys. 60:83–92. 51. Sra, J.S., M.R. Jazayeri, B. Avitall, A. Dhala, S. Deshpande, Z. Blanck, 48. Liu, J.E., R.T. Hahn, K.M. Stein, S.M. Markowitz, P.M. Okin, R.B. De- and M. Akhtar. 1993. Comparison of cardiac pacing with drug therapy in the vereux, and B.B. Lerman. 1996. Does mechanoreceptor stimulation explain treatment of neurocardiogenic (vasovagal) syncope with bradycardia or asys- neurally mediated syncope? Circulation. 94(Suppl. I):I-334. tole. N. Engl. J. Med. 328:1085–1090. 49. Kapoor, W.N., M.A. Smith, and N.L. Miller. 1994. Upright tilt testing in

2744 Mosqueda-Garcia et al.